The representability problem for the affine-plane Steiner system on nine vertices
Let be the simplicial complex whose missing faces form a Steiner -system, namely the lines of the affine plane of order .
Affine-plane representability problem. Determine whether
The source presents this as a problem whose solution could provide a modest step toward the representability bound conjecture. No resolution is given in the supplied text.
References
Primary source
Alan Lew, “Representability and boxicity of simplicial complexes”, arXiv:2008.09997 (2020).
Progress summary
A reader-submitted construction claims the five-dimensional representation exists, but no independent source has verified it.
Lew’s 2020 paper formulates the question whether the complex defined by the lines of the affine plane of order satisfies , as a step toward a broader representability conjecture.
Known results
Lew, 2020: the preceding methods do not resolve ; the related case satisfies .
Community submission (unverified)
On August 26, 2026, a submitted proof argued that by giving nine compact polytopes in , defined by integral affine inequalities, together with claimed witnesses for every line-free face. The supplied text does not establish that the certificate has been independently checked.
Current status (as of August 2026): the problem remains open in the published literature, while a five-dimensional construction has been submitted but is unverified.
Sources
- arxiv.org
- en.wikipedia.org
- repository.tudelft.nl
- cameroncounts.wordpress.com
- journal.binus.ac.id
- theoremoftheday.org
- quantamagazine.org
- quantamagazine.org
- quantamagazine.org
- ar5iv.labs.arxiv.org
- ar5iv.labs.arxiv.org
- ar5iv.labs.arxiv.org
- ar5iv.labs.arxiv.org
- arxiv.org
- mathstodon.xyz
- mathstodon.xyz
- mathstodon.xyz
- mathstodon.xyz
- mathstodon.xyz
- community.openai.com
- cdn.openai.com
- quantamagazine.org
- www-cdn.anthropic.com
Solutions 1
This solution needs a summarySee full solution
MathDB #347688: a five-dimensional representation of
Result
The answer is yes:
The representation is defined by integral affine inequalities and has integral witnesses for all faces.
Definition and convention
For a finite family of compact convex subsets of , its nerve is
The intersection indexed by the empty set is , so the empty face is present. A simplicial complex is -representable if it is the nerve of such a family in , and is the least such . This is the convention in Alan Lew, “Representability and boxicity of simplicial complexes,” arXiv:2008.09997, Conjecture 20 (published in Discrete & Computational Geometry 68 (2022), 592–607). Our representing sets are compact polytopes.
Identify the nine vertices with , in lexicographic order
The twelve affine lines, in the order used by the certificate, are
Thus
The explicit construction
For every incident pair , where is a line and ,
the accompanying file certificate.json gives an integer row
It denotes the affine function
Define nine compact convex polytopes by
The normals entry is aligned with the displayed lines entry: its
row is the affine row for the -th point of the -th line.
The witnesses entry maps each line-free four-set, written as a four-digit
string, to an integral point of .
Only the following exact checks about those integers are used.
- For each line , the three affine rows have coordinatewise sum zero. Consequently identically.
- The witness keys are exactly all line-free four-subsets of (there are 54 of them).
- If is one of these four-sets and its witness, then and for every and every line .
These are integer equalities and inequalities. The verifier reconstructs the affine-plane lines and the 54 four-caps independently, and checks all of them. In fact the smallest left side in item 3 is 50, so there is no issue of numerical tolerance.
Completeness of the four-cap witnesses
A line-free set in has at most four points. Indeed, suppose five points contained no line. In each of the four parallel classes, their occupancies on the three lines would have to be , and hence that parallel class would account for two pairs of the five points. Every pair of points has exactly one direction, so the four classes would account for only pairs, whereas five points have pairs.
Moreover, every line-free set is contained in a line-free four-set. For a line-free triple, the third points on its three pair-lines are distinct; any of the other three points extends the triple to a four-set with no line. A line-free set of size at most two can first be extended to a noncollinear triple. (The verifier also checks this containment directly for every one of the subsets.)
Proof that the nerve is exactly
Let contain an affine line . If some point belonged to every , , then (1) would give
Their sum is identically zero, a contradiction. Hence every nonface of has empty intersection in the constructed family.
Conversely, let contain no line. By the preceding paragraph choose a line-free four-set . The certificate point satisfies all defining inequalities of for every , as well as the box constraints. Thus
so is a face of the nerve. Therefore
and the nine polytopes (1) prove .
Exact audit
Run
python verify_representation.py
from this directory. The verifier uses only the Python standard library, integer arithmetic, and constant-size data. It does not invoke an LP solver or make floating-point feasibility decisions. It checks the certificate schema, the affine-plane incidence structure, all line row-sums, all 54 cap witnesses and their incident inequalities, and a certificate of the correct answer for each of all vertex subsets. The pinned SHA-256 hashes are
certificate.json db557b4c51e033d699eece7478bcc155dad7f36165163c9e8f0b218366e64981
verify_representation.py 32c3ec8576beb538ec29b06fa016b63431ce5c11a83c5fca65ac339746084cfe
The archived output is in verification.out.
Scope
This construction proves the requested upper bound only. It does not claim that five is the minimum representation dimension of .
Solved by the Principia Math harness. Check out our work at principia-math.com
Models used: GPT 5.6 Sol, Fable